denovoadp segfaults, better tool?

[k-gun12]

I'm trying to convert a SOLiD assembly to dna space. I ran solid_denovo_postprocessor.pl on the afg file generated from the assembly, and now I've got a set of colorspace reads again. The final line of velvet output gave me the "Final graph has etc etc ,max 137561," so I'm assuming that's the length I'm supposed to feed to denovoadp? When I do so:

denovoadp ./postprocessor-output 137561

It reads a few sequences, gets to >NODE_1051-0 and then segfaults. I toyed with increasing the max length, but it always segfaults when it hits that read. If I increase it enough it dumps core:
*** glibc detected *** denovoadp: double free or corruption (out): 0x00007faedc9ba010 ***

Is there another tool to convert the postproc output to dnaspace??

[k-gun12]

Also, I can't help but sound off for a second on the sheer nightmare that is working with SOLiD data. ABI's woeful "pipeline" is nothing but a mish mash of 3rd party, poorly documented perl scripts in conjunction with velvet. We didn't make the platform choice ourselves.. but goodness, I don't see how they ever even planned to stay competitive in the sequencer market with such poor support.

[kopi-o]

For sure, the documentation for the Velvet-related scripts is very poor ... they could at least try to give you an idea about what the file formats look like!

[Torst]

Also, I can't help but sound off for a second on the sheer nightmare that is working with SOLiD data. ABI's woeful "pipeline" is nothing but a mish mash of 3rd party, poorly documented perl scripts in conjunction with velvet.

We are mainly Illumina and 454 users. One colleague got some "free" SOLiD runs done. For the life of me, I was never able to get them de novo assembled with Velvet! I think I was doing things correctly, but perhaps the 24bp read length was the issue in the end. On the plus side, we successfully mapped the SOLiD reads to a reference using SHRiMP rmapper-cs. But the poor documentation was enough to put me off forever.

[KevinLam]

I'm trying to convert a SOLiD assembly to dna space. I ran solid_denovo_postprocessor.pl on the afg file generated from the assembly, and now I've got a set of colorspace reads again. The final line of velvet output gave me the "Final graph has etc etc ,max 137561," so I'm assuming that's the length I'm supposed to feed to denovoadp? When I do so:

denovoadp ./postprocessor-output 137561

It reads a few sequences, gets to >NODE_1051-0 and then segfaults. I toyed with increasing the max length, but it always segfaults when it hits that read. If I increase it enough it dumps core:
*** glibc detected *** denovoadp: double free or corruption (out): 0x00007faedc9ba010 ***

Is there another tool to convert the postproc output to dnaspace??

I totally agree with you. Their documentation reminds of the default README and INSTALL that come with compiling GNU software.

anyway have you tried removing that read and running it?
alternatively extract that read and run on that alone?

[Blaize]

Denovo ADP fault

I've got the same "Segmentation fault" error msg. at the last step when converting to basespace, trying at least five samples... (fragment, and mate pair)
There's one and only alternative to do some de novo assembly with Solid data: CLC Genomics Workbench as I know at this time. It has two feature (normal and beta 3) about this kind of assembly.

RGRDS: Balázs

[zee]

I got the same error too. Funny enough when I compiled it on my Mac it worked perfectly.

My Mac specs are:

x86_64-apple-darwin10.0
i686-apple-darwin10-g++-4.2.1 (GCC) 4.2.1 (Apple Inc. build 5646) (dot 1)
Copyright (C) 2007 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

Program:
denovoadp: Mach-O 64-bit executable x86_64

[zee]

Hi Everybody,

Attached is a fix for denovoadp, thanks to Colin Hercus at Novocraft (www.novocraft.com).

Just unzip and place denovoadp.cxx with this new file and run

make clean && make.

I will try to get this fix back to ABI solidsoftwaretools.com.

[orcy]

For what its worth, I ran this software fine on out 64 bit linux server. Theres a really simple answer if your having that much of a problem with the available software. write your own.

[zee]

OK, so the patch actually does fix a logic error dealing with color space reads that contain "." characters in either of the first 2 bases of the read.
If you're data does have this characteristic then you wont have any problems. However, it's likely to arise in many cases even after you have run `saet` or some other analysis filter.